upgrading from 3.0.1_GA_160 to latest on FC4

[Mike Carifio]

Any gottas I should be aware of before I start?

I've looked at http://wiki.zimbra.com/index.php?tit...up_Procedu re
for backing up my existing installation. I've installed in /opt/zimbra.
Will recursive copy of the directory to (say) /opt/zimbra_3.0.1_GA_160
also copy the database stuff?

Pl. advise. Thanks.

[jholder]

Nope.
Looks good. We'll be here if you need any help.

-john

[Mike Carifio]

I've attached the installation log.

[root@postmaster zimbra]# dig $(hostname); dig usys.com mx

; <<>> DiG 9.3.1 <<>> postmaster.usys.com
;; global options: printcmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 26808
;; flags: qr aa rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 0, ADDITIONAL: 0

;; QUESTION SECTION:
;postmaster.usys.com. IN A

;; ANSWER SECTION:
postmaster.usys.com. 3600 IN A 192.168.111.5

;; Query time: 1 msec
;; SERVER: 192.168.111.4#53(192.168.111.4)
;; WHEN: Wed Feb 21 17:18:00 2007
;; MSG SIZE rcvd: 53


; <<>> DiG 9.3.1 <<>> usys.com mx
;; global options: printcmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 39363
;; flags: qr aa rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 0, ADDITIONAL: 1

;; QUESTION SECTION:
;usys.com. IN MX

;; ANSWER SECTION:
usys.com. 3600 IN MX 10 postmaster.usys.com.

;; ADDITIONAL SECTION:
postmaster.usys.com. 3600 IN A 192.168.111.5

;; Query time: 1 msec
;; SERVER: 192.168.111.4#53(192.168.111.4)
;; WHEN: Wed Feb 21 17:18:01 2007
;; MSG SIZE rcvd: 69

[root@postmaster zimbra]#

[Mike Carifio]

To see if the installation has a dns issue, I did a new install on another machine.
This went without a hitch. So there's something wrong on the server machine itself.

[phoenix]

The obvious question is: is the FQDN of your Zimbra server actually postmaster.usys.com? It seems to not exists as far as DNS is concerned:

Creating server entry for postmaster.usys.com...
*** Running as zimbra user: /opt/zimbra/bin/zmprov -l cs postmaster.usys.com

ERROR: service.FAILURE (system failure: createServer) (cause: javax.naming.NamingException [LDAP: error code 80 - internal error])
Done

Setting spell check URL...
*** Running as zimbra user: /opt/zimbra/bin/zmprov -l ms postmaster.usys.com zimbraSpellCheckURL http://postmaster.usys.com:7780/aspell.php

ERROR: service.FAILURE (system failure: unable to lookup server by name: postmaster.usys.com message: [LDAP: error code 80 - internal error]) (cause: javax.naming.NamingException [LDAP: error code 80 - internal error])
Done

You need to check you /etc/hosts file for correct entries, you also need to fix the DNS entries. Search the forums for 'DNS in a Nutshell' and follow the instructions there (it's also in the wiki). If you're behind a NAT device then you probably need a split-DNS setup to resolve the server IP correctly - details also in the wiki for that.

This check on your DNS shows no server by the name postmaster.usys.com.

[Mike Carifio]

I run a local dns server that forwards to my ISP for local names it can't
resolve. This is what you mean by a "split server"?

Since I have an operational zimbra 4.5.1 on a different box (I pointed
the mx record there and installed), can I tar up /opt/zimbra/db, carry it
to the new box and untar it in place? The 3.0 to 4.5.1 upgrade appears
to have correctly updated all my mailboxes (well, no errors), so swapping one for the
other should be ok? So I would stop zimbra (/sbin/service zimbra stop),
tar up the old /opt/zimbra/db (tar zcf /tmp/zimbra-db.tgz /opt/zimbra/db)
rename the old db directory (mv /opt/zimbra/db /opt/zimbra/db.old), untar
the updated stuff into /opt/zimbra/db and restart the server. Or is there
a preferred (better) way to do this?

--------------------

I think my machine is postmaster, see below.

[root@postmaster zimbra.broken]# ifconfig eth0
eth0 Link encap:Ethernet HWaddr 00:0D:88:36:43:C6
inet addr:192.168.111.5 Bcast:192.168.111.255 Mask:255.255.255.0
inet6 addr: fe80::20d:88ff:fe36:43c6/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:52551 errors:0 dropped:0 overruns:0 frame:0
TX packets:16029 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:5136516 (4.8 MiB) TX bytes:2095577 (1.9 MiB)
Interrupt:5 Base address:0xb400

[root@postmaster zimbra.broken]# grep postmaster /etc/hosts
192.168.111.5 postmaster.usys.com postmaster

[root@postmaster zimbra.broken]# hostname
postmaster.usys.com



[root@postmaster zimbra.broken]# dig postmaster.usys.com

; <<>> DiG 9.3.1 <<>> postmaster.usys.com
;; global options: printcmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 49479
;; flags: qr aa rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 0, ADDITIONAL: 0

;; QUESTION SECTION:
;postmaster.usys.com. IN A

;; ANSWER SECTION:
postmaster.usys.com. 3600 IN A 192.168.111.5

;; Query time: 1 msec
;; SERVER: 192.168.111.4#53(192.168.111.4)
;; WHEN: Thu Feb 22 09:55:37 2007
;; MSG SIZE rcvd: 53

[phoenix]

Split DNS is detailed in the wiki here. You should also disable ipv6 on the Zimbra server.

You posted the install logs earlier, they show this error:

Creating server entry for postmaster.usys.com...
*** Running as zimbra user: /opt/zimbra/bin/zmprov -l cs postmaster.usys.com

ERROR: service.FAILURE (system failure: createServer) (cause: javax.naming.NamingException [LDAP: error code 80 - internal error])
Done

Setting spell check URL...
*** Running as zimbra user: /opt/zimbra/bin/zmprov -l ms postmaster.usys.com zimbraSpellCheckURL http://postmaster.usys.com:7780/aspell.php

ERROR: service.FAILURE (system failure: unable to lookup server by name: postmaster.usys.com message: [LDAP: error code 80 - internal error]) (cause: javax.naming.NamingException [LDAP: error code 80 - internal error])
Done

Specifically this line:

ERROR: service.FAILURE (system failure: unable to lookup server by name:

that seems self-explanatory to me. Your DNS isn't set-up correctly. It can't be found, you need A & MX records pointing to this server - the install routine doesn't seem to think one or other of them exists.

[Mike Carifio]

The command line dig queries show all is well (doesn't it)?

I set up a "split DNS" locally according to the instructions on the wiki.
I reinstalled 4.5.1, I get the same result.

Here's what dig says:

$ dig postmaster.usys.com
; <<>> DiG 9.3.1 <<>> postmaster.usys.com
;; global options: printcmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 63369
;; flags: qr aa rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 1, ADDITIONAL: 0

;; QUESTION SECTION:
;postmaster.usys.com. IN A

;; ANSWER SECTION:
postmaster.usys.com. 2592000 IN A 192.168.111.5

;; AUTHORITY SECTION:
postmaster.usys.com. 2592000 IN NS 192.168.111.5.postmaster.usys.com.

;; Query time: 1 msec
;; SERVER: 192.168.111.5#53(192.168.111.5)
;; WHEN: Thu Feb 22 11:15:52 2007
;; MSG SIZE rcvd: 81




[root@postmaster sysconfig]# dig postmaster.usys.com mx

; <<>> DiG 9.3.1 <<>> postmaster.usys.com mx
;; global options: printcmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 60266
;; flags: qr aa rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 1, ADDITIONAL: 1

;; QUESTION SECTION:
;postmaster.usys.com. IN MX

;; ANSWER SECTION:
postmaster.usys.com. 2592000 IN MX 10 postmaster.usys.com.

;; AUTHORITY SECTION:
postmaster.usys.com. 2592000 IN NS 192.168.111.5.postmaster.usys.com.

;; ADDITIONAL SECTION:
postmaster.usys.com. 2592000 IN A 192.168.111.5

;; Query time: 1 msec
;; SERVER: 192.168.111.5#53(192.168.111.5)
;; WHEN: Thu Feb 22 11:19:09 2007
;; MSG SIZE rcvd: 97

[phoenix]

No, the dig command doesn't show all is well. You're doing a dig on the FQDN of your server, it should be on the domain name only i.e. dig usys.com any - does that show your postmaster.usys.com as the mail server with all the correct records?

[Mike Carifio]

Dig and usys.com are in sync now. Reinstalled 4.5.1, got the same
name resolution errors. I think I'm at the stage where I want to do
a clean install, but I don't want to lose my existing mailboxes. Is
there a way to set the mailboxes and mail aside, install (not upgrade),
and then restore the mailboxes? Pl. advise.

Thanks for your help.

---------------------------------------

[root@postmaster zcs]# dig usys.com any

; <<>> DiG 9.3.1 <<>> usys.com any
;; global options: printcmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 6560
;; flags: qr aa rd ra; QUERY: 1, ANSWER: 3, AUTHORITY: 0, ADDITIONAL: 1

;; QUESTION SECTION:
;usys.com. IN ANY

;; ANSWER SECTION:
usys.com. 2592000 IN SOA usys.com. postmaster.usys.com. 10118 43200 3600 3600000 2592000
usys.com. 2592000 IN NS postmaster.usys.com.
usys.com. 2592000 IN MX 10 postmaster.usys.com.

;; ADDITIONAL SECTION:
postmaster.usys.com. 2592000 IN A 192.168.111.5

;; Query time: 32 msec
;; SERVER: 192.168.111.5#53(192.168.111.5)
;; WHEN: Thu Feb 22 12:50:59 2007
;; MSG SIZE rcvd: 119